LFM2-700M: benchmark results
Liquid AI's LFM2-700M, one of four LFM2 checkpoints from 350M to 2.6B parameters, built with a hybrid architecture of multiplicative gates and short convolutions for edge deployment. Provider: Liquid AI. Released 2025-07-10. Access: Open.
Aggregate Rating 1237 ± 19, rank #1551 of 1654 rated models, from 351 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval Polish Summarization - PSC | 24.94 | chrF3++ (%) | 90.5 |
| Enkrypt AI - Bias Risk | 73.64 | Risk Score | 74.7 |
| EuroEval Spanish Summarization - Mlsum ES | 26.94 | chrF3++ (%) | 69.4 |
| EuroEval Portuguese Summarization - Publico | 31.6 | chrF3++ (%) | 55 |
| EuroEval Serbian Summarization - LR SUM SR | 25.89 | chrF3++ (%) | 51.7 |
| EuroEval Albanian Summarization - LR SUM SQ | 28.83 | chrF3++ (%) | 50.3 |
| Kurakurai French LLM - GPQA Diamond | 25.8 | Accuracy (%) | 48.8 |
| EuroEval Italian Summarization - Ilpost SUM | 35.89 | chrF3++ (%) | 48 |
| EuroEval French Summarization - Orange SUM | 34.96 | chrF3++ (%) | 47.3 |
| Kurakurai French LLM - IFEval | 53.79 | Accuracy (%) | 46.3 |
| Enkrypt AI - Toxicity Risk | 3.79 | Risk Score | 45.9 |
| EuroEval German Summarization - Mlsum DE | 31.09 | chrF3++ (%) | 42.9 |
Interactive version: theaggregate.ai/model?slug=lfm2-700m · How It Works · Data refreshed daily, snapshot 2026-10-24.